As data became the new oil with the emergence of Deep Learning, storing and collecting data has been an expensive process. Due to the voluminous amount of data being added to the internet each day, our systems should have the capability to adapt to the changes in the new data while remembering the knowledge from the past. Continual learning can be made use of to give systems the capability of remembering, and dealing with past knowledge, while learning from data continuously generated throughout the world.

The event commenced with a welcome and introduction about the speaker, Mr. Jathushan Rajasegaran given by the Staff Advisor of the IEEE SPS Student Branch Chapter of University of Moratuwa. The speaker proceeded with the webinar by intuitively explaining basics on continual learning and few-shot learning with the use of concepts found in novel research papers from this field. Later he explained how above learning methods could be unified via meta-learning. The session was conducted in a very interactive and interesting manner. Furthermore, the attendees were benefited with a 30 minute Q&A session to clarify any subject matter from the webinar session. The event concluded with the Vote of Thanks by the Secretary of the society.

About the speaker

Mr. Jathushan Rajasegaran

University of California Berkeley, California, USA


Mr. Jathushan Rajasegaran received his BSc in Electronics and Telecommunication Engineering from the University of Moratuwa, Sri Lanka in 2018. Following his undergraduate, he spent 2 years at the Inception Institute of Artificial Intelligence, UAE, working on continual learning, meta-learning, and few-shot learning. After that, he joined the University of California Berkeley, and he is currently a first-year Ph.D. student at BAIR. His research interests are mainly in learning methods for deep neural networks including continual learning, meta-learning, and few-shot learning, and 3D vision.